摘要
文章提出一种基于YOLOv5模型结合数据增强的印章与签字检测方法。为解决实际场景中进行印章和签字标注人工成本较高的问题,提出了一种印章签名的数据增强方法,并结合深度学习算法YOLOv5进行建模,实现了小数据样本下的印章与签字检测建模。与传统的检测方法相比,该方法可以有效地提高检测准确率和精确率,并且具有更好的鲁棒性,可以应用于电子政务等业务系统中。
The article proposes a seal and signature detection method based on YOLOv5 model combined with data augmentation.To solve the problem of high labor cost of performing seal and signature labeling in practical scenarios,a data augmentation method for seal and signature is proposed and combined with the deep learning algorithm YOLOv5 for modeling to realize the seal and signature detection modeling under small data samples.Compared with traditional detection methods,this method can effectively improve detection accuracy and precision and has better robustness,which can be applied in business systems such as e-government.
作者
李欣宸
金子轩
赵一寒
LI Xinchen;JIN Zixuan;ZHAO Yihan(Beijing-Dublin International College,Beijing University of Technology,Beijing 100124,China)
出处
《信息与电脑》
2023年第4期194-197,共4页
Information & Computer
基金
国家级大学生创新创业训练计划项目“基于YOLOv5模型的印章识别方法及工具”(项目编号:GJDC-2022-01-66)。
